ankita_dev
ServiceNow Employee

Overview: 

The Knowledge Graph (KG) application enhances consuming features—portal searches, Now Assist Virtual Agent, AI agents, and generative AI skills—by surfacing contextual data from ServiceNow records like knowledge bases, and external sources. By creating a semantic layer that connects structured and unstructured information, it enables more intelligent, relevant, and personalized responses across the platform. 

 

ServiceNow Knowledge Graph application facilitates the capabilities below: 

  1. Natural language queries:  Allow users to query structured ServiceNow data conversationally—for example, “Who is my manager?”; “What’s the status of my incident?” etc.  
  2. User context: Provide Now Assist Virtual Agent and AI Agents with contextual information such as role, department, or location to deliver personalized and relevant responses. 
  3. Slot-filling: Simplify user interactions by automatically populating known information in forms or chat requests, reducing effort and friction. 
  4. Smarter AI Agents: Empower agents to retrieve insights, facts, and relationships directly from the knowledge graph to automate actions and improve accuracy. 
  5. Integration with Workflow Data Fabric: Use an integrated data layer that gives real-time, secure, and unified access to data from any source without duplicating it, enabling faster, data-rich workflows across the platform. 

For more details on the Knowledge Graph, refer to the links below: 

Virtual Agent Properties for Knowledge Graph enablement 

To enable Knowledge Graph for Now Assist Virtual Agent, ensure below properties are set accordingly:  

  • sn_vad_genai.knowledge_graph.enabled  set to true 
  • sn_ais_assist.enable_knowledge_graph_nlq  set to true 
  • sn_ais_assist.disable_knowledge_graph_nlq_on_portal set to false 
  • sn_ais_assist.kgnlq_schema_name: Controls the information available via Knowledge Graph in Now Assist in Virtual Agent. The default value is “sn_kg.now_user_graph_nlq”. 

For more details on Virtual Agent system properties, refer to this link. 

Note: In Dec’25 release, the Enterprise Graph option is hidden by default for Now Assist Virtual Agent and Now Assist Panel, and only Enterprise Graph (Small) can be selected as schema option. To use Enterprise Graph, Admins can enable it as a schema option by setting the property “sn_kg.va.enable_global_graph_all_modes” to true. For more information, please visit the below links: 

Configuring custom Knowledge Graphs to meet specific customer needs 

Often, customers may want to configure KG to display additional custom fields as part of KG or hide certain fields that they perceive as sensitive or for other reasons, which otherwise are included by default. For example, out of the box, the KG user schema displayed as part of Virtual Agent and Portal search results includes details like user title, location, direct reports, and manager details. However, many customers may consider such data as sensitive and wish to remove it from the KG data displayed. Such situations need to be handled by modifying the schema using the KG Designer application.  

Refer to the guidance below using above example as a use case to help configure this.  

1. Copy the Knowledge Graph schema using the guidance provided here. 

2. Note the copied Graph name and edit the KG as per below steps:

a. Depending on the customization needed, you can either select existing nodes to update, add new nodes to the Knowledge Graph Schema by clicking the “Add Nodes” button, or delete nodes by selecting a particular node and clicking the “delete” option in the right pane. 

ankita_dev_0-1769599189293.png

b. Once you add or select a Node, you can add or update the following details in the navigation pane 

  • Node details 
  • Columns that can be queried 
  • Related Nodes 

ankita_dev_0-1769599287101.png

c. In the Node details section, you can edit the Node description and add node synonym. 

ankita_dev_1-1769599324586.png

d. In the Columns that can be queried section, search for and select any desired columns or unselect currently selected columns to remove them from the search results and click on Save. 

ankita_dev_2-1769599379616.png

e. Add, delete or edit edges in the Related nodes section if needed and select Save. 

ankita_dev_3-1769599454513.png

 3. Create a new context configuration. 

a. Navigate to the table “sn_kg_context_config” which lists the Knowledge Graph Context Configuration records. 

ankita_dev_4-1769599526239.png

b. Create a new record with the following details, such as: 

  • Context name → user_context_new 
  • Graph → new Graph you copied and customized 
  • Configuration → Copy the settings from the existing user_context record 

ankita_dev_5-1769599572707.png

c. Save the newly created record. 

4. Update the system properties below: 

a. sn_kg.kg.context.default_user_context → Set to your new context name (e.g., user_context_new) 

b. sn_kg.kg.context.default_user_context_schema → Set to your copied graph name 

c. sn_ais_assist.kgnlq_schema_name → Set to your copied graph name 

 5. Enable Knowledge Graph in the Assistant. 

a. Navigate to Conversational Interfaces > Assistants 

b. Click and navigate to the assistant that is used in the Portal. 

c. Navigate to Knowledge Graphs > Set the value as the copied Knowledge Graph for "Slot filling" and "Natural Language query". For more details, refer this link.

ankita_dev_6-1769599619283.png

d. Click ‘Save and continue’. 

6. Test and verify that the system now uses the new user context and the created Knowledge Graph. 

1 Comment